“Hay Niemand” wake word model for ESP32 and ESPHome
microWakeWord model for the wake word “Hay Niemand”, trained on microwakeword.com from synthetic speech (English (GB), English (US), German). It runs on-device on ESP32-S3 boards with ESPHome's micro_wake_word component, including the Home Assistant Voice Preview Edition, with no cloud involved.
Model facts
- Wake word
- “Hay Niemand”
- Language of the training voices
- English (GB), English (US), German
- Trained
- 2026-07-30 · microWakeWord v2
- Training steps
- 20,000
- Recall, with background noise
- 63.8%
- False activations on the benchmark set
- 1.51 per hour at probability cutoff 0.5
- Sliding window
- 5
- File
- quantized INT8 TFLite, 59 KB, with a micro_wake_word manifest
- Downloads
- 0
Add it to your ESPHome device
ESPHome fetches the model straight from this site when it compiles the firmware (ESPHome 2024.7 or newer); nothing to download.
micro_wake_word:
models:
- model: https://microwakeword.com/api/models/1493/manifest.jsonSign in on this page to download the ZIP (TFLite + manifest) for offline use. Models are licensed for personal and non-commercial use by default (commercial license).
Similar wake words in the library
- “Alexander” — 61.8% recall
- “Okay Rock” — 47.6% recall
- “sweetie darling” — 39.6% recall
- “Natasha” — 3.6% recall
- “brother home” — 97.8% recall
- “Hey Odin” — 77.0% recall
Custom wake words are trained in the Training Center; the library lists every public model with the same measurements.